Provenance of far-traveled nappes in the eastern Mediterranean (Uppermost Unit, Crete): constraints from U–Pb zircon ages of detrital and igneous zircons

Abstract

New U–Pb ages of detrital and igneous zircons of the Uppermost Unit of Crete shed light on its provenance and on Eohellenic to Eoalpine imprints in the eastern Mediterranean. The detrital zircons of all nappes show Variscan ages and are characterized by a Minoan-type age spectrum, which is typical for the NE margin of Gondwana. Parts of the metasedimentary rocks are unexpectedly young. Their detrital zircon ages continue via the Permian until the Late Triassic, Middle Jurassic and Early Cretaceous. The high-grade metamorphic rocks of the Asterousia crystalline complex are likely equivalents of the low-grade metamorphic trench and fore-arc deposits of the Vatos nappe pointing to Late Cretaceous slab roll back. The presence of both late Permian detrital zircons and Late Cretaceous arc-type granitoids suggest that the Uppermost Unit of Crete is derived from the late Permian/Late Cretaceous magmatic belt situated north of the Sava–Vardar–Izmir–Ankara Suture in the Strandja–Rhodope area. To achieve their recent position on Crete, the nappes had to travel more than 500 km. The traveling path is well tracked by rocks of the Upper Cycladic Unit, which are similar to those of the Uppermost Unit of Crete. The large displacement of the Cretan nappes was controlled not only by nappe transport, but probably also by dextral strike–slip along the North Anatolian Fault Zone and related counterclockwise rotation of the Anatolian block since the Eocene.

Graphical abstract

Introduction

The rocks of the eastern Mediterranean realm show evidence for Grenvillian, Cadomian, Variscan, Cimmerian, and Alpine orogenic events. During each of these events, large volumes of felsic igneous rocks developed, which commonly include magmatic and inherited zircons. After weathering and decomposition, the zircons were removed from the rock, transported and deposited in various basins as detrital zircons. Analysis of the type, composition and age of these detrital zircons help to reconstruct the positions of the different plates and nappes through time. Concerning the nappe stack of Crete, such type of provenance analysis is particularly important, as metamorphism and deformation related to the orogenic events damaged fossils and affected some of the isotopic systems, which are relevant for constraining the protolith ages of metasedimentary and metaigneous rocks. Moreover, the primary position of volcanic and sedimentary rocks was affected by thrusting and nappe emplacement due to subduction/collision and by extensional collapse of thickened crust. Consequently, there are fragments of lithosphere, the database of which is still incomplete and our understanding of its geological evolution is poor. Such fragments are present in the nappes of the Uppermost Unit of Crete (Fig. 1), which is the most enigmatic nappe stack of the eastern Mediterranean commonly attributed to the Internal Hellenides.

Fig. 1

Schematic sketch of the Cretan nappe stack and map of the Aegean region (modified from Chatzaras et al., 2016) showing areas of the Attico-Cycladic Massif in dark grey and exposures of Uppermost Unit in red. Exposures of Uppermost Unit after Bonneau (1972, 1984), Seidel et al. (1976, 1981), Dürr (1985), Dürr et al. (1978), Krahl et al. (1982), Reinecke et al. (1982), Thorbecke (1987), Altherr et al. (1994), Patzak et al. (1994), and Langosch et al. (2000). (BFZ, Bornova Flysch Zone; LN, Lycian Nappes; MM, Menderes Massif; SZ, Sakarya Zone; SMRM, Serbo-Macedonian and Rhodope Massifs

In contrast to the Uppermost Unit, the lower nappes of Crete (Plattenkalk, Phyllite–Quartzite s.str., Pre-Cimmerian Basement, and Tyros Unit), are part of the External Hellenides (e.g., Robertson 2012; Chatzaras et al. 2012). Detrital zircons separated from the lower nappes were investigated during several studies focusing on the closure of the Paleotethys. These studies revealed (1) a Minoan-type age spectrum characterized by a large number of Tonian/Stenian detrital zircons and a lack or only very few Mesoproterozoic zircons, apart from Stenian (Keay and Lister 2002; Zulauf et al. 2007, 2015; Marsellos et al. 2012; Dörr et al. 2015), and (2) Variscan to Cimmerian docking of microplates along the southern margin of Laurasia (Chatzaras et al. 2012; Zulauf et al. 2015, 2018, 2019). Moreover, the rocks deposited along the northern margin of Gondwana are largely free of Cambrian granite and Paleozoic detrital zircons, whereas those deposited along the southern active margin of Laurasia show evidence for early Paleozoic and Carboniferous to early Permian magmatism.

While the closure of the Paleotethys is relatively well constrained, there is still little consensus on the regional plate tectonic development during the Eohellenic (Tithonian/Early Cretaceous) and Eoalpine (Late Cretaceous) orogenic events. As no generally accepted geodynamic model exists for this period, examination of the regional geology based on field evidence is required. Defining the original provenance of the individual nappes of the Uppermost Unit of Crete remains the major tool in unraveling their movement and emplacement within the present nappe stack. Although paleomagnetic studies can be used to show how the continental separation between Gondwana and Eurasia has evolved through time (e.g., van Hinsbergen and Schmid 2012), there are still different opinions concerning the location and movement of intervening microplates. The most disputable topics are (1) the location and polarity of subduction of the Vardar lithosphere (Fig. 1), (2) the position of the individual nappe complexes of the Internal Hellenides, and (3) the provenance, movement and emplacement of those rocks, which are today forming the nappes of the Uppermost Unit.

In the present contribution, we present new U–Pb ages of zircons, which have been separated from metasediments and metaigneous rocks of the Uppermost Unit of Crete (Fig. 2a). As in some of the investigated rocks, the temperature during Eohellenic and Eoalpine metamorphism was < 400 °C, the protolith age of sediments can be determined using fossils. Moreover, the cooling history of most of the investigated rocks is well constrained by K–Ar and ³⁹Ar–⁴⁰Ar ages of hornblende and white mica, and by fission-track ages of apatite. The stratigraphy of the individual nappes of the Uppermost Unit (Bonneau and Lys 1978; Krahl et al. 1982; Zulauf et al. 2023a) provides the basis for a reasonable interpretation of the detrital zircon ages and of geological processes through time.

Fig. 2

a Map of Crete showing the different geological units (modified after Zulauf et al. 2023a, and references therein). Sample localities are indicated. b Nappe stacks of the Uppermost Unit resting on top of the Pindos Unit of Crete. Yellow stars indicate nappes from which samples were collected for the present study. Key to references: (1) Reinecke et al. (1982), (2) Seidel et al. (1976), (3) Seidel et al. (1977), (4) Martha et al. (2017), (5) Martha et al. (2019), (6) Liati et al. (2004), (7) Bonneau and Lys (1978), (8) Zulauf et al. (2023a), (9) Koepke et al. (2002), (10) Malten (2019), (11) Koepke (1986), (12) Koepke et al. (1997), (13) Tortorici et al. (2012), (14) Zulauf et al. in prep., (15) Bonneau et al. (1974), (16) Robert and Bonneau (1982), (17) Palamakumbura et al. (2013), and (18) Karakitzios (1988)

We will present (a) new ages of detrital zircons of metasedimentary rocks, which belong the Uppermost Unit of Crete, (b) new ages of zircons of a granitic pebble of the Pindos flysch, and (c) new ages of zircons of felsic igneous veins, which cut through high-grade metasedimentary rocks of the Uppermost Unit. The new ages, together with the shape and internal structure of the zircons, and the deformation microfabrics and composition of other detrital grains (quartz and feldspar), are used to reveal provenance areas for late Paleozoic and Mesozoic times and to reconstruct the distribution of lithospheric plates in the Eastern Mediterranean. The new ages also help to constrain the Eohellenic to Eoalpine orogenic activity within the Internal Hellenides. For simplicity, the emplacement directions of nappes or the direction of subduction are mainly referred to here in terms of modern geographical coordinates. The 2022 GSA Geologic Time Scale (http://www.geosociety.org/science/timescale/timescl.pdf) is used throughout.

Regional geology

The Uppermost Unit of Crete consists of several nappes, which are resting on top of the Pindos or on deeper units (Fig. 2a). There are two different stacks (Fig. 2b), which are briefly described in the following passages. One of these nappe stacks is exposed in central Crete in the area of Vatos–Preveli–Spili–Gerakari, near Gonies, west of Kamares, and on the island of Gavdos. This stack consists of the Preveli, Vatos, Ophiolite and Mourne nappes (Zulauf et al. 2023a, and references therein; Fig. 2a, Fig. 2c Supplement).

The Preveli nappe consists of Permian metasiliciclastic rocks covered by limy phyllite and middle to late Permian limestone (now marble) (Bonneau and Lys 1978), and bimodal metavolcanic rocks. The metavolcanic rocks developed in a Triassic extensional setting and led to the formation of skarn (Zulauf et al. 2023a). Towards the stratigraphic top, the mafic metavolcanic rocks dominate and are related to Upper Triassic to? Jurassic (meta)chert and cherty marble (Krahl et al. 1982). The Preveli rocks underwent Eohellenic deformation under epidote-blueschist-facies conditions at T = ca. 370 °C (Koepke et al. 1997; Tortorici et al. 2012; Zulauf et al. 2023b). Apatite fission-track ages of phyllite of the Preveli nappe are 30 ± 2 Ma (Thomson et al. 1999). The main tectonic transport during late nappe emplacement was top-to-the west.

The Vatos nappe (Aubouin and Dercourt 1965) consists of wildflysch-like Upper Jurassic to Upper Cretaceous metasediments and volcanics (Bonneau et al. 1977; Krahl et al. 1982; Malten 2019). Calcschists include Upper Cretaceous fossils and display a synkinematic crystallization of white mica, calcite, quartz and albite along the main foliation (Koepke 1986; Tortorici et al. 2012). Deformation microfabrics of quartz and calcite and the metamorphic index minerals, such as actinolite, chlorite, albite and white mica indicate lower greenschist-facies metamorphism (Karakitzios 1988; Malten 2019), which must be younger than Uppermost Cretaceous (the age of the fossils) but older than the Palaeocene–Eocene Pindos flysch underneath. Of particular interest are limy metaconglomerates, which include rodingite and other ophiolitic pebbles (Krahl et al. 1982) and Late Jurassic/Early Cretaceous foraminifera (Malten 2019). U–Pb dating of calcite, which precipitated in pressure shadows behind rigid clasts of the limy metaconglomerate, yielded 76 ± 4 Ma (Malten 2019), which is probably the age of the greenschist-facies metamorphism.

The Ophiolite nappe includes deformed metaultramafic rocks (serpentinite, metahornblendite) and meta-pillowbasalt (Krahl et al. 1982). A detailed petrographic description of the serpentinites, which are forming the basal part, is given in Koepke (1986) and Tortorici et al. (2012). Relics of spinel lherzolite result from undepleted mantle evolved along a spreading ridge with low spreading rate and deformed at very high temperature (Koepke et al. 2002). The serpentinite is overlain by metahornblendite, which was overprinted under greenschist-facies conditions. U–Pb SHRIMP dating of zircons of the metahornblendite yielded 163 ± 3 Ma, which is interpreted as crystallization age (Liati et al. 2004). K–Ar ages of hornblende yielded ca. 156 Ma (Koepke et al. 2002) meaning that the hornblendite was formed close to the Middle/Late Jurassic boundary and emplaced and cooled down to T < 500 °C still during the Late Jurassic. Gabbroic/dioritic veins, which cut through the serpentinite, yielded K–Ar hornblende ages at ca. 140 Ma, interpreted as cooling after melt emplacement (Koepke et al. 2002). The gabbroic veins were overprinted under greenschist-facies conditions at ca. 95 Ma. Younger veins of quartzdiorite and tonalite, which cut through the rocks mentioned above, are almost undeformed. One metahornblendite yielded an apatite fission-track age at 35 ± 6 Ma (Thomson et al. 1999).

The Mourne nappe, on top of the Ophiolite nappe, is exposed west of Mourne and between Kerames and Korifi (Fig. 2c Supplement). It consists of blueschist, amphibolite, marble and micaschist/gneiss. The blueschists contain relics of omphacite. Amphibole and phengite, separated from the crystalline rocks of Gavdos and Gerakari, yielded similar K–Ar ages at ca. 148 Ma (Seidel et al. 1977), which are compatible with the fact that the blueschists of Gavdos are unconformably overlain by Late Cretaceous (meta)sediments (Vicente 1970; Seidel et al. 1977). Thus, similar to the Ophiolite nappe, the Mourne nappe should have been situated at upper structural levels (T < ca. 350 °C, the closure T of the K–Ar isotopic system of white mica) since the Upper Jurassic, while the Vatos Unit and the rocks of the second nappe stack (see below) underwent latest Cretaceous/Paleocene metamorphism under various conditions.

The second nappe stack consists of four subnappes, which are (from bottom to top) Arvi Unit, Miamou Unit, Greenschist Unit and Asterousia Crystalline Complex (ACC) (Fig. 2d–f Supplement). The Arvi Unit (Davi and Bonneau 1972; Bonneau 1973) consists of terrigenous turbidites, prehnite–pumpellyite facies metamorphic pillow basalts and reddish (meta)limestones. Maastrichtian–?Palaeocene microfossils are present in red calcareous mudstone matrix of associated lava breccia (Robert and Bonneau 1982; Papanikolaou 1989), that should have been part of an Upper Cretaceous seamount in the northern domain of the Pindos ocean (Palamakumbura et al. 2013). The Miamou Unit consists of flyschoid terrigenous sediments, diabase and limestones, the latter of late Jurassic age (Bonneau et al. 1974). These rocks underwent greenschist-facies metamorphism (Bonneau et al. 1974). The Greenschist Unit results from basic volcanic rocks with a MORB-type signature (Reinecke et al. 1982) and metasedimentary rocks. Paleocene greenschist-facies top-to-the SE displacement is documented not only on Anafi (Martha et al. 2016), but also in the Melambes and Pefkos areas of Crete (Martha et al. 2017, 2019).

The Asterousia Crystalline Complex (ACC) consists of high-grade metamorphic rocks (largely metasediments) and metagranitoids (Bonneau 1972; Davi and Bonneau 1972; Seidel et al. 1976, 1981; Langosch et al. 2000; Martha et al. 2019), which reflect a Late Cretaceous magmatic arc. Apart from Crete, ACC-type rocks are further present in the Upper Cycladic Unit such as Anafi, Donoussa, Nikouria, Syros, and Tinos (Fig. 1). Conditions of the low-pressure/high-temperature metamorphism of the ACC of Crete have been constrained at P = 400–500 MPa and T_max = 700 °C (Seidel et al. 1981). Slices of serpentinite have been incorporated into the metasedimentary rocks prior to amphibolite facies metamorphism (Bonneau 1973; Reinecke et al. 1982; Be’eri-Shlevin et al. 2009), while a large number of I-type granitoids intruded the metasedimentary successions of Anafi, western Asterousia, Melambes, and Pachia Amos during a short period in the middle to late Campanian (78–72 Ma; U–Pb on zircon, Kneuker et al. 2015; Martha et al. 2016, 2017, 2019). K–Ar biotite and hornblende ages of Cretan ACC-type rocks range from 76 to 66 Ma, and apatite fission-track ages of the ACC of Lentas range from 31 to 20 Ma (Thomson et al. 1998).

Analytical techniques

Laser ablation-inductively coupled plasma-mass spectroscopy

Samples of metasedimentary rocks collected for isotope mass spectrometry were processed at the Institut für Geowissenschaften of Frankfurt University using standard mineral separation techniques. These include crushing and sieving, before concentration of heavy minerals by heavy liquids (bromoform, methyleniodide) and magnetic-separation with a Frantz isodynamic separator. Hand-picked detrital zircon grains were mounted in 25-mm-diameter circular epoxy mounts and polished to expose a section at their inner core. Prior to their analysis, the grains were examined using cathodoluminescence (CL) imaging in order to reveal their internal structure and to identify cracks and mineral inclusions. U–Pb isotope analysis of detrital zircons was performed by laser ablation-inductively coupled plasma-mass spectroscopy (LA-ICPMS) using a Thermo-Finnigan Element II sector field ICPMS attached to a New Wave LUV213 laser ablation system (λ = 213 nm). Ablation was carried out in a He carrier gas in a low volume (2.5 cm³) cell; laser beam parameters used were 30 μm diameter; 5 Hz repetition rate 75% power output. Isotope data were acquired in peak-jumping mode. Background and ablation data for each analysis were collected over 90 s, with background measurements (carrier gas, no ablation) being taken over the first 30 s prior to initiation of ablation. Data were collected at time-resolved mode allowing acquisition of the signal as a function of time (ablation depth), and subsequently recognition of isotopic heterogeneities within the ablated volume. Raw data were processed offline using an Excel^® spreadsheet program created by A. Gerdes. Mass discrimination of the MS, and elemental fractionation during laser ablation were corrected by calibration against the GJ-1 zircon standard (Jackson et al. 2004), which was analyzed routinely during analytical sessions (three standard analysis at the beginning and end of every session of 33 unknowns, and 2 standard analyses every 10 unknowns). Prior to this correction, the change of elemental fractionation (e.g., Pb/U and Pb/Th ratios as function of ablation time and thus depth) was corrected for each set of isotope ratios by applying a linear regression through all measured ratios versus time, excluding some outliers (> 2 s.e.), and taking the intercept t = 0 as the correct ratio. Changes in isotopic ratios arising from laser drilling into domains of distinct Pb/U ratio (core/rim), mineral inclusions, and zones affected by Pb loss (metamictization/cracks), can usually be detected by careful monitoring of the time-resolved signal, such analyses are normally rejected. Common Pb correction was applied only when the interference- and background-corrected ²⁰⁴Pb signal was significantly higher than the detection limit of about 20 cps. The latter is limited by the amount of Hg in the carrier gas and the accuracy to which the ²⁰²Hg and thus the interfering ²⁰⁴Hg can be monitored. Corrections made were based on common Pb composition given by the second stage growth curve of Stacey and Kramers (1975) for Neoproterozoic age (600 Ma). Data presentation was made with Isoplot (Ludwig 1994).

Isotope dilution thermal ionization mass spectrometry

Zircon grains with the size of 100–200 μm were hand-picked from the > 1.6 Amp fraction of the heavy mineral separates for the isotope dilution thermal ionization mass spectrometry (ID-TIMS). As the ID-TIMS ages of zircons are more precise than the ages obtained using LA-ICPMS, this method was applied for determining emplacement ages for the igneous rocks. After washing with 6 N HCl and acetone, zircon grains were weighted, transferred to small Savillex beakers and put with 24 N HF for 6 h on the hotplate at 90 °C. The HF was decanted and the vials were loaded again with 24 N HF and a mixed ²⁰⁵Pb/²³⁵U spike. The small Savillex vials were then placed into a Parr bomb. After dissolution at 180 °C for 96 h and subsequent evaporation to dryness at ca. 80 °C on a hotplate, the sample was converted into chloride by adding 0.2 ml 3 N HCl. Chemical separation of Pb and U on 100 μl columns (ion-exchange resin AG 1 × 8, 100–200 mesh) followed the method of Krogh (1973). The U and Pb isotope ratios of zircon were obtained using a Finnigan MAT 261 mass spectrometer in static multicollector mode with simultaneous ion counting of ²⁰⁴Pb. All isotopic ratios were corrected for mass fractionation (1.0 ± 0.3‰/a.m.u), blank (ca. 5 pg) and initial lead using the Stacey and Kramers (1975) model Pb composition. The Pb/U isotope ratios were plotted using Isoplot (Ludwig 2001), with error ellipses reflecting 2σ uncertainty. To verify the complete analytical procedure, standard zircons 91,500, Gj 1 and M257 have been reproduced (see data of the Institut für Geowissenschaften of Frankfurt University in Nasdala et al. 2008).

Microfabrics

All investigated samples underwent deformation and metamorphism. Thin sections from the samples were investigated to reconstruct the kinematics, the deformation–crystallization relations, and the deformation mechanisms of quartz and feldspar. In addition to metamorphic index minerals, the latter were used to estimate the metamorphic conditions.

Results

Locality, petrography and microfabrics of studied samples

Metasedimentary rocks

Sample 180922-1 is a fine-grained siliciclastic metaconglomerate of the basal part of the Preveli nappe, exposed 800 m SW of Lefkogia in central Crete (Fig. 2a, Fig. 2c Supplement, Table 1 Supplement). The rock is well foliated. Displaced quartz veins indicate a top-to-the SE sense of shear. Main constituents are quartz, plagioclase, K-feldspar, white mica, chlorite and opaque phases (Figs. 3a, 4a). Quartz is present in form of clasts and as newly grown fibers in pressure shadows behind rigid clasts. The clasts show evidence for both ductile and brittle deformation. Ductile deformation is weak and documented by undulatory extinction and subgrains. Serrated grain boundaries are present in both rigid clasts and newly grown quartz fibers. The quartz clasts, on the other hand, show various types of inherited fabrics including recrystallized fabrics with strikingly different grain size and chessboard-like subgrains. Exsolution of perthite lamellae in K-feldspar is also regarded as an inherited fabric (Fig. 4a).

Fig. 3

Field photographs of studied rocks. a Siliciclastic metaconglomerate of the Preveli nappe exposed SW of Lefkogia (sample 180922–1); view is subparallel to the WNW–ESE trending stretching lineation; length of hammer = 28 cm. b Vatos wildflysch exposed WSW of Ardaktos showing quartzite layer at the top, from which sample HM10 was collected. c Detail of quartzite (sample HM10); length of compass = 9 cm. d Discordant folded granitic vein (sample 0909–17-1) in quartzite of the Asterousia Crystalline Complex (ACC) is cut by normal faults; road cut between Krotos and Lentas. e Close-up view of contact between fold hinge of granitic vein and greyish foliated quartzite (sample Do552); same locality like in (d). f Boudinaged granitic vein (sample Ld002) in metasediments of the ACC; road cut between Krotos and Lentas

Fig. 4

Microphotographs of investigated rocks. a Metaconglomerate of basal Preveli nappe with large amount of detrital feldspar (180922–1, SW of Lefkogia). b Quartzite of Vatos flysch showing evidence for dissolution–precipitation creep in form of dark stylolites (HM 10, W of Ardaktos). c High-temperature deformation of granitic vein is indicated by chessboard-like subgrains in quartz, formation of myrmekite (red arrow), and sutured phase boundaries (080917–1, road cut between Lentas and Krotos). d Red granite of Pindos flysch (Do525, W of Melambes). Ductile deformation is indicated by serrated grain boundaries of quartz and deformation twins in plagioclase

Sample HM10 is a quartzite of the Vatos flysch exposed W of Ardaktos in central Crete (Fig. 2a, c Supplement, Table 1 Supplement). The quartzite forms the Uppermost part of this flysch sequence (Fig. 3b) and displays primary bedding in form of light and dark layers (Fig. 3c). The bedding plane dips moderately towards NNE. The quartzite is cut by a subhorizontal brittle shear plane, which shows a N–S trending striation. The quartzite results from badly sorted sandstone. Apart from quartz, there are other components such as white mica, plagioclase, chlorite and opaque phases. The large quartz grains show evidence for deformation portrayed by serrated grain boundaries and stylolites (Fig. 4b), which indicate both dislocation and dissolution–precipitation creep as the main deformation mechanisms. Inherited fabrics are similar to those described from sample 180922-1.

Sample Mel11 is made of quartzite of the Greenschist Unit exposed 1.5 km SW of Nea Krya Vrysi (Melambes area, central Crete, Fig. 2a, d Supplement, Table 1 Supplement). The quartzite is impure and consists of ca. 75% of quartz and minor amounts of garnet, K-feldspar, plagioclase, white mica, clay minerals and accessories. Quartz shows a very fine-grained dynamically recrystallized fabric. Relics of small, elongated older grains display undulatory extinction and subgrains. Similar to the micas, most of the quartz grains show a weak shape-preferred orientation. Garnet was largely replaced by chlorite. Further details concerning the content and texture of this quartzite are described in Martha et al. (2017).

Sample Mel51 consists of quartzite collected from the ACC 1 km SSE of Nea Krya Vrysi (Melambes area, central Crete, Fig. 2a, d Supplement, Table 1 Supplement). The quartzite consists of quartz, K-feldspar, plagioclase, garnet, biotite, epidote, white mica and chlorite. The fabric of quartzite is mylonitic with ribbons of quartz and mineral clasts in a fine-grained, dynamically recrystallized matrix consisting mostly of quartz. A striking foliation results from almost pure layers of recrystallized quartz in alternation with impure layers. For further details of this rock, see Martha et al. (2017).

Sample Pf038 is a quartzite of the ACC exposed in Sykologos village (Pefkos area of eastern Crete, Fig. 2a, f Supplement, Table 1 Supplement). Quartzite is impure and consists of > 75% of quartz, with diopside, biotite, white mica, clay minerals and opaque phases as accessories. Quartz shows an equilibrium fabric (foam structure) of polygonal and small grains with few larger elongated grains forming clusters or thin layers. The grain size of quartz is correlated with the amount of impurities, quartz grains being larger in almost pure layers. The shape-preferred orientation of quartz in quartzite is supported by the grain-shape fabrics of biotite and clay minerals. For further details, see Martha et al. (2019).

Sample Pf040 consists of chiastolite hornfels of the ACC, which has been collected 200 m SE of Ag. Georgios (Pefkos area of eastern Crete, Fig. 2a, f Supplement, Table 1 Supplement). The chiastolite hornfels consists of quartz, biotite, plagioclase, andalusite, K-feldspar, white mica, garnet, chlorite, clay minerals and graphite (Martha et al. 2019). Quartz is present in form of fine-grained recrystallized matrix and aggregates of large recrystallized grains. Quartz shows weakly developed chessboard patterns based on prism- and basal-parallel subgrains. Biotite inclusions and pinning of quartz grain boundaries by biotite are common. K-feldspar shows undulatory extinction, while Karlsbad twinning is rare. Garnet may occur as euhedral porphyroblasts or is anhedral in layers. Garnet porphyroblasts are strongly disintegrated and were partly replaced by white mica, chlorite and clay minerals. For further details of this rock, see Martha et al. (2019).

Sample Do552 is a quartzite of the ACC exposed between Krotos and Lentas (central Crete, Fig. 2a, e Supplement, Table 1 Supplement). The grey quartzite is well foliated (Fig. 3e) and intercalated with biotite schist and paragneiss. Quartz of the quartzite displays a pervasively recrystallized fabric. This sequence is cut by granitic veins described below (Fig. 3d–f).

Meta-igneous rocks

Sample 080917–1 was collected from a granitic dike, which cuts through metasediments of the ACC between Krotos and Lentas (Fig. 2a, e Supplement, Table 1 Supplement). The thickness of the vein varies from 50 cm to ca. 2 m. Macroscopic deformation of the vein is indicated by open to tight folds (Fig. 3e) and by normal faults (Fig. 3d). Constituent minerals of the vein are quartz, K-feldspar, plagioclase and minor biotite, which is partly replaced by chlorite. Although the vein is free from foliation, it shows evidence for ductile deformation. The original quartz grains were up to 1 cm in diameter. Ductile deformation of quartz was accommodated by recrystallization driven by strain-induced grain boundary migration, which led to grain-size reduction. High-temperature deformation is indicated by chessboard-type subgrains in quartz and by serrated phase boundaries between quartz and K-feldspar (Fig. 4c). K-feldspar is present in form of microcline, which is up to 2 cm in diameter and shows flame perthite and replacement by myrmekite. K-feldspar is partly recrystallized along discrete planes, which show newly grown tiny crystals. Plagioclase is up to 5 cm in diameter. Some crystals show primary zonation and inclusions of white mica.

Sample Ld002 was collected from a granitic sill, which is aligned subparallel to the foliation of the ACC metasediments cropping out between Krotos and Lentas (Fig. 2a, e Supplement, Table 1 Supplement). The sill is similar in composition and deformation microfabrics like the dike described above from sample 080917-1 but includes a large number of titanite, which is up to 300 µm in size. Moreover, the macroscopic deformation is portrayed by boudinage, which reflects extension subparallel to the foliation of the wall rock (Fig. 3f).

Sample Do525 was collected from a reddish granite pebble that is several meters in diameter and part of the Pindos flysch exposed along a field path W of Melambes (Fig. 2a, d Supplement, Table 1 Supplement). The granite consists of quartz, plagioclase, K-feldspar and chlorite. Quartz is up to 8 mm in diameter. Apart from young fractures, there is evidence for earlier high-temperature deformation in form of serrated grain boundaries and chessboard-type subgrains in quartz (Fig. 4d). Plagioclase is up to 4 mm in diameter and shows striking deformation twins after the albite and pericline law. The twin lamellae are often bent or kinked. K-feldspar is up to 6 mm in diameter and affected by sericitization. Both plagioclase and K-feldspar are reddish and partly replaced by calcite, which is affected by a large number of type 2 twins (after Burkhard 1993). Twinned calcite is also present along fractures. The serrated phase boundaries between quartz and plagioclase and between plagioclase and K-feldspar are further evidence for high-temperature deformation. Chlorite is up to 4 mm long and seems to have developed by replacement of former biotite.

U–Pb zircon data

Metasedimentary rocks

Siliciclastic metaconglomerate of Preveli Unit (sample 180922-1)

Fifty-four zircons have been separated from the Preveli metaconglomerate sample 180922-1. The zircons vary significantly in size and in shape. The largest zircon is 250 µm in diameter. Most zircons are well rounded to subrounded and display a striking oscillatory zonation in CL images (Fig. 5a). There are up to three growth stages. Some of the zircons are pitted or fractured, and only few are euhedral in shape.

Fig. 5

CL images of selected zircon grains

Only few zircons are older than Ediacaran (Fig. 6a, b, Table 2 Supplement). They include Tonian/Stenian as well as Paleoproterozoic and Neoarchean ages. Eight zircons are Ediacaran and 13 zircons are Cambrian–Ordovician in age. The rest of the zircons are Carboniferous with peaks at 351 ± 3 Ma, 337 ± 2 Ma and 319 ± 3 Ma. The youngest zircon is 305 ± 6 Ma.

Fig. 6

Density plots of detrital zircons of Permian siliciclastic metaconglomerate of the Preveli nappe (a, b), and of quartzite of the Vatos flysch (c, d). Note the different scales of the horizontal axes

Quartzite of Vatos flysch (sample HM10)

Fifty zircons have been separated from the quartzite sample HM10. The size of the zircons varies from 40 to 200 µm. The shape and structure of the zircons is similar to those described from sample 180922-1. Most zircons are well rounded to subrounded and display a striking oscillatory zonation in CL images (Fig. 5b). The number of fractured zircons is high, and only few are euhedral.

Nine zircons are Paleoproterozoic in age (Fig. 6c, d, Table 3 Supplement). Apart from Stenian types, Mesoproterozoic zircons are lacking. Four zircons are Tonian, four are Cryogenian, and 5 are Ediacaran in age. The number of early Paleozoic (Cambrian–Ordovician) zircons is very low. Most of the zircons are Carboniferous to early Permian, with peaks at 320 ± 3 Ma, 305 ± 2 Ma and 292 ± 4 Ma (Fig. 6c). The youngest zircon is 289 ± 6 Ma.

Quartzite of greenschist unit (sample Mel11)

Seventy-four zircons have been separated from the quartzite sample Mel11. The size of the zircons varies from 70 to 250 µm. About a quarter of the zircons is euhedral with slightly rounded edges. The other zircons are subrounded to well rounded. Most of the zircons display a striking oscillatory zoning (Fig. 5c). Some zircons are fractured. The zoning of the zircons reflects up to three growth stages. Few zircons show holes in the central part or along the margin.

Only 15% of the zircons are older than 1 Ga, two of which are Archean, seven are Neoproterozoic and further two are Mesoproterozoic (Fig. 7a,b, Table 4 Supplement). Eleven zircons are Ediacaran with peaks at 575 ± 5 Ma and 544 ± 6 Ma (Fig. 7a). Cambrian zircons are lacking and two are Ordovician in age. Devonian and early Carboniferous zircons are also scarce. Most zircon ages are straddling the Carboniferous/Permian boundary with a small peak at 303 ± 3 Ma and the most important peak at 291 ± 2 Ma (Fig. 7a). The middle and late Permian period is also reflected by the zircon ages with a peak at 273 ± 2 Ma. Triassic zircons are also present with a Norian peak at 214 ± 2 Ma. The youngest detrital zircon is Late Jurassic (157 ± 4 Ma, Table 4 Supplement).

Fig. 7

Density plots of detrital zircons of quartzites of Greenschist Unit and ACC collected from the area of Melambes (a, b, e, f), Lentas-Krotos (c, d), and Pevkos (g, h). Note the different scales of the horizontal axes

Quartzites of Asterousia Crystalline Complex

The preparation of sample Do552 of ACC quartzite yielded 102 detrital zircons. The size of the zircons varies from 40 to 260 µm. Most zircons are rounded to subrounded and display a striking oscillatory zoning with up to three growth stages (Fig. 5g). Some of the zircons are pitted. A large number of zircons is present as fragments, which results from fracturing of former intact zircons.

The number of zircons > 1 Ga is low (24%, Fig. 7d, Table 5 Supplement). Most of these old zircons are Paleoproterozoic with a peak at 1865 ± 11 Ma. There is only one Mesoproterozoic zircon, while the number of Tonian zircons is high at 10, with a peak at 987 ± 14 Ma (Fig. 7d). Zircons with ages from Cryogenian to Ordovician are almost lacking. The number of Silurian zircons is elevated with a peak at 422 ± 6 Ma (Fig. 7c). The same holds for Carboniferous zircons, which display peaks at 355 ± 5 and 303 ± 3 Ma. There is an almost complete and continuous spectrum of Permian and Triassic zircons with peaks at 280 ± 3 Ma, 253 ± 2 Ma (main peak), 240 ± 3 Ma, and 225 ± 2 Ma. The youngest zircon grain is late Norian at 207 ± 6 Ma.

Seventy-two zircons have been separated from the quartzite sample Mel51. The zircons are similar to those described above from sample Mel11. Their size varies from 50 to 270 µm. Euhedral zircons with rounded edges are frequent. Some show sector zoning. Half of the zircons are well rounded. Most of these rounded zircons are also broken and some are pitted (Fig. 5d). The oscillatory zoning of the zircons reflects up to two growth stages.

Two zircons are Archean and 8 zircons are Paleoproterozoic (Fig. 7f, Table 6 Supplement). Four zircons are Tonian. The number of Ediacaran zircons is six with a peak at 578 ± 5 Ma (Fig. 7e). There are also Cambrian and Ordovician zircons with a peak at 483 ± 4 Ma. The number of Silurian, Devonian and Carboniferous zircons is very low. Many zircons are Permian in age, with the most important peak at 291 ± 3 Ma and a further moderate peak at 272 ± 3 Ma (Fig. 7e). Ten zircons are Triassic with a Norian peak at 213 ± 2 Ma. Six zircons are Jurassic with a Hettangian peak at 198 ± 2 Ma (Fig. 7e). The two youngest detrital zircons are Middle Jurassic (170 ± 4 Ma) and Late Cretaceous (Cenomanian, 98 ± 2 Ma).

The preparation of sample Pf038 of ACC quartzite yielded 89 detrital zircons. The size of these zircons varies from 70 to 260 µm. About half of the zircons are euhedral with slightly rounded edges (Fig. 5e). Based on their oscillatory zoning, up to two growth stages are indicated. Sector zoning is present but not common. Few zircons show striking holes. The other zircons are subrounded to well rounded. The zoning of these zircons reflects up to four growth stages.

The number of zircons older than 1 Ga is low at 18% (Fig. 7h, Table 7 Supplement). Three zircons are Archean and eleven are Paleoproterozoic. Nine zircons are Tonian with a peak at 724 ± 14 Ma. The number of Ediacaran and Cambrian zircons is low, whereas that of Ordovician and Silurian is higher with peaks at 464 ± 6 Ma and 427 ± 5 Ma, respectively (Fig. 7g). The number of Carboniferous zircons is high with a moderate peak at 336 ± 4 Ma and the most important peak at 322 ± 3 Ma. Zircons with ages throughout the entire Permian period are also present with peaks at 293 ± 6 Ma and 258 + 5/−10 Ma (Fig. 7g). Eight zircons are Triassic with a Rhaetian peak at 204 ± 3 Ma. This is the youngest age of detrital zircons of this sample.

Chiastolite hornfels of Asterousia Crystalline Complex

Preparation of the chiastolite bearing hornfels exposed near Pefkos (sample Pf040) yielded only 16 zircons, four of which with bad quality concerning age dating (concordancy < 90%, Table 8 Supplement). The size of the zircons varies from 100 to 200 µm. Only few of the zircons are euhedral and well preserved. They show oscillatory zoning. The other zircons are subrounded and are present as pitted fragments of former larger intact crystals (Fig. 5f).

As many granitoids of the ACC intruded at 74–75 Ma (Kneuker et al. 2015; Martha et al. 2017), and the chiastolite hornfels is clearly of contact metamorphic origin related to a non-exposed pluton, the youngest concordant zircons at 74 ± 2 Ma and 75 ± 2 Ma are interpreted as the age of contact metamorphism (Fig. 8a; Martha et al. 2019). The zircons, which are older than 75 Ma, should be of detrital origin. Apart from one Tonian zircon, there are Silurian and Devonian zircons with a peak at 408 ± 14 Ma (Fig. 8a). Carboniferous zircons are lacking, whereas Permian zircons are present at 270 ± 8 Ma, 262 ± 8 Ma, and 253 ± 4 Ma (Fig. 8a, Table 8 Supplement). The youngest detrital zircon with a reliable age is Middle Jurassic (167 ± 5 Ma). There is one zircon with an age at 82 ± 3 Ma, which, however, displays a low degree of concordance (84%, Table 8 Supplement).

Fig. 8

U–Pb concordia plot of analyzed zircons of chiastolite hornfels of ACC (a), of granite boulder in Pindos flysch (b), and of granitoid veins of ACC (c, d)

Igneous rocks

Reddish granite boulder in Pindos flysch

Preparation of the reddish granite boulder (sample Do525) yielded 30 zircons, which are largely euhedral. The size of the zircons is up to 180 µm. Most of the zircons show up to 7 holes. Some are strongly resorbed along the margins, while other zircons are fractured. Based on the oscillatory zoning, two growth stages are inferred. Two of the zircons are two small and could not be used for age dating. The seven youngest zircons yielded a concordia age at 234 ± 2 Ma, which is interpreted as the emplacement age of the granite (Fig. 8b). Inherited zircons yielded ages at 245 ± 6, 246 ± 6, 262 ± 7 and 271 ± 7 Ma (Table 9 Supplement).

Granitoid veins in Asterousia Crystalline Complex

Six euhedral zircons have been separated from a granitoid dike, which cuts through quartzite of the ACC near Lentas (sample 080917-1). These zircons are similar to those of the granitoid sill described below. Their size varies from 100 to 350 µm, and only one growth stage is documented. The zircons yielded a concordia age at 72.8 ± 0.2 Ma (Fig. 8c, Table 10 Supplement). This age is interpreted as emplacement age of the melt.

Preparation of a granitoid sill (sample Ld002) yielded 70 zircons. The size of the zircons varies from 50 to 300 µm. All of the zircons are euhedral, but in most cases fractured (Fig. 5h). Four of the zircons are of bad quality concerning age dating and thus were not considered (Table 11 Supplement). The other zircons yielded a concordia age at 72.5 ± 0.4 Ma (Fig. 8d), which should reflect the time of melt emplacement.

Discussion

Provenance of the Vatos and Preveli Unit

Inherited deformation microfabrics of detrital grains, such as quartz or feldspar, are well preserved in the Preveli metaconglomerate and in the Vatos quartzite, because the temperature during metamorphism was low (< 400 °C, see references above). The inherited fabrics of detrital quartz and feldspar grains of the Preveli metaconglomerate and of the Vatos quartzite unequivocally indicate high-grade metamorphic rocks in the source areas. This holds particularly for the exsolution of perthite lamellae in K-feldspar and for the chessboard pattern in quartz. This type of subgrains results from the combination of prism- and basal-parallel subgrains, which develop at T > 600 °C in the stability field of high-quartz (Mainprice et al. 1986; Kruhl 1996).

The depositional age of the Preveli metaconglomerate is constrained to the early/middle Permian based on the age of the youngest detrital zircon (latest Carboniferous, 305 ± 6 Ma) and the middle to late Permian age of the Preveli marble on top (Bonneau and Lys 1978). Because of the large amount of Carboniferous detrital zircons and the high-grade metamorphic inherited fabrics of detrital quartz and feldspar, the Preveli metaconglomerate and related siliciclastic rocks are regarded as Variscan molasse. The entire detrital zircon age spectrum of the Preveli metaconglomerate, with a clear dominance of Carboniferous zircons, few Silurian and Cambro-Ordovician zircons, an Ediacaran age peak, the presence of Tonian/Stenian zircons and a Mesoproterozoic age gap (Fig. 9), is similar to the detrital zircon age spectrum of the pre-Cimmerian basement and of the Tyros Unit of the lower nappes, which are part of the External Hellenides (Zulauf et al. 2015, 2018; Chatzaras et al. 2012).

Fig. 9

Compilation of new U–Pb ages of magmatic, inherited, metamorphic and detrital zircons of the Uppermost Unit and of Pindos flysch. ACC = Asterousia Crystalline Complex

The Ediacaran detrital zircons should be derived from felsic magmatic rocks, which emplaced during the assembly of northeastern Gondwana (e.g., Abdelsalam et al. 2002; Kröner and Stern 2005; Küster et al. 2008). Tonian/Stenian zircons and a Mesoproterozoic age gap are characteristic for northeast Gondwana and related Minoan Terranes (Avigad et al. 2003; Zulauf et al. 2007, 2015; Küster et al. 2008; Stern et al. 2010; Meinhold et al. 2013; Dörr et al. 2015). This holds particularly for the Arabian–Nubian shield (Meinhold et al. 2021), which largely comprises Neoproterozoic juvenile oceanic island arcs (e.g., Stern et al. 2010; Johnson et al. 2011). This Minoan-type age spectrum is characteristic not only for the pre-Cimmerian basement of the External Hellenides, but has also been detected in the Cyclades (e.g., Keay and Lister 2002; Löwen et al. 2015; Hinsken et al. 2015; Bröcker et al. 2016; Seman et al. 2017), western Turkey (Ustaömer et al. 2012a, b, 2016; Zlatkin et al. 2012), the Serbo-Macedonian Massif (Meinhold et al. 2010), the basement of the Pannonian Basin (Pozsgai et al. 2016), Sicily (Williams et al. 2012), Sardinia (Avigad et al. 2012), and the eastern Alps (Heinrichs et al. 2012) (see also age compilation in Fig. 13 of Dörr and Stein 2019).

The Cambrian and Ordovician detrital zircons of the Preveli metaconglomerate are probably related to igneous activity along a late Cadomian magmatic arc situated at the margin of northeastern Gondwana. Associated (meta)granitoids are today exposed in the pre-Cimmerian basement of eastern Crete (Romano et al. 2004) and on the Peloponnesus (Dörr et al. 2015). The separation of the Cambrian magmatic arc from northeastern Gondwana started with the Devonian opening of the Paleotethys (Stampfli et al. 2013). Few of the early Paleozoic detrital zircons could also be derived from the Arabian–Nubian shield, where related magmatic rocks are also exposed (Fritz et al. 2013).

Based on the youngest detrital zircon of the Vatos quartzite (289 ± 6 Ma), the maximum deposition age of this rock is middle Permian and thus significantly older than the Jurassic/Early Cretaceous maximum deposition age of the Vatos beds as constrained by faunal composition (Malten 2019). For this reason, our field-based assumption that the sedimentary sequence of the quartzite is part of an olistolith within the Vatos flysch, is corroborated. As the number of inherited high-grade metamorphic fabrics in the Vatos quartzite is similarly high as in the Preveli metaconglomerate, we interpret the detrital components of the Vatos quartzite to be derived also from Variscan high-grade metamorphic and igneous rocks.

As the protoliths of the Vatos Unit are consistent with the main constituents of the ACC-type rocks (see below), we regard both as equivalents, meaning that the protoliths of all these rocks were deposited in a Late Jurassic to Late Cretaceous trench to fore-arc setting and accreted to the upper plate during progressive subduction. Because of slab roll-back, the arc was ‘shifted’ towards the accreted Vatos-type rocks, which explains their HT–LP metamorphic overprint and the intrusion of calc-alkaline plutons and veins in Campanian times.

The trench and fore-arc deposits mentioned above were deposited, while the rocks of the Preveli Unit underwent Eohellenic (Late Jurassic to Early Cretaceous) subduction and related HP–LT metamorphism (Seidel et al. 1977; Koepke et al. 1997; Zulauf et al. 2023b). The pre-Permian detrital zircons of the Preveli Unit are similar to those of the ACC and Greenschist Unit, which are characterized by Carboniferous (Variscan) ages and a Minoan-type age spectrum. A correlation of the Preveli blueschists with HP–LT rocks of Greek mainland and of adjacent domains (Seidel et al. 1977) is supported by findings of late Jurassic (Tithonian) blueschists in the Vardar Zone (Robertson 2012, and references therein), in the Pelagonian Zone (Altherr et al. 2023), and in the Strandja Massif and Circum–Rhodope Belt (Fig. 10; e.g., Okay et al. 2001; Sunal et al. 2011; Burg 2012, and references therein; Okay and Nikishin 2015; Liati et al. 2016). Moreover, the Triassic rift-related volcanic rocks of the Preveli nappe (Zulauf et al. 2023a) might form equivalents of the Triassic A-type granitoids of the Serbo-Macedonian Zone (Himmerkus et al. 2009), which also show evidence for a Eohellenic HP–LT metamorphism (Mposkos et al. 2021, and references therein).

Fig. 10

Distinct rock types of Eastern Mediterranean, which are critical for the provenance of the Uppermost Unit of Crete: (1) Triassic volcanic rocks and granitoids, (Bröcker and Keasling 2006; Bröcker and Pidgeon 2007; Himmerkus et al. 2006; Bulle et al. 2010; Akal et al. 2011; Chatzaras et al. 2012; Liati et al. 2013; Zulauf et al. 2013, 2023a; Bonev et al. 2019; this study); (2) Mesozoic ophiolites (Robertson 2002, and references therein; Liati et al. 2004; Smith 2006, and references therein; Koglin et al. 2009a; Robertson et al. 2012, and references therein); (3) Late Jurassic/early Cretaceous HP–LT metamorphic rocks of the Internal Hellenides (after Seidel et al. 1976, 1977; Michard et al. 1994; Wawrzenitz and Mposkos 1997; Most et al. 2001; Most 2003; Kydonakis et al. 2015; Liati et al. 2016; Altherr et al. 2023); (4) Late Jurassic and Early Cretaceous HP–LT metamorphic rocks of the CPS (Aygül et al. 2016, and references therein); (5) Upper Cretaceous (arc-type) magmatic rocks (von Quadt et al. 2005, 2007, and references therein; Okay and Nikishin 2015, and references therein; Kneuker et al. 2015; Martha et al. 2016, 2017). (6) Permian igneous rocks (Reischmann 1998; Liati 2005; Koglin et al. 2009b, Tanatsiev et al. 2012; Okay and Nikishin 2015, and references therein; Antic et al. 2015, 2016; Bonev et al. 2019; Lazarova et al. 2021; Salacinska et al. 2021). CPS = Central Pontides Supercomplex; DSFZ = Death Sea fault zone; EASZ = East Anatolian fault zone; IPS = Intrapontide suture; IAES = Izmir–Ankara–Erzincan suture; NAFZ = North Anatolian fault zone; S = Samothraki ophiolite

Besides the possible source areas in northern Greece, Balkan and NW Turkey, the Preveli and Vatos rocks might have been derived from the Central Pontide Supercomplex (Fig. 10; Okay et al. 2006; Aygül et al. 2016), where Early Cretaceous eclogites and blueschists are exposed (Altherr et al. 2004; Okay et al. 2006). Towards the south, a further metabasite–marble-dominated unit, characterized by Middle Jurassic blueschist-facies metamorphism, is present (Aygül et al. 2016). Possible equivalents of the Vatos rocks are exposed in the Eselner and Dady Unit (Okay et al. 2013).

The detrital zircon isotopic data of the Istanbul Zone and of the Proto-Pelagonian metasediments, on the other hand, include Mesoproterozoic zircons and a lack of Neoproterozoic zircons, pointing to an Amazonian source for these sequences (Ustaömer et al. 2011; Zlatkin et al. 2014). These areas can be ruled out as possible source for the rocks of the present study.

Late Cretaceous arc-type granitoids and provenance of the Asterousia Crystalline Complex

Besides the late Permian detrital zircons, the Late Cretaceous arc-type granitoids and their host rocks are important tracers to unravel the provenance of the Greenschist Unit and of the ACC of Crete. Cretaceous HT–LP metamorphic rocks and Campanian I-type (meta)granitoids are main constituents of the ACC (Seidel et al. 1976, 1981; Langosch et al. 2000; Kneuker et al. 2015; Martha et al. 2017, 2019). The new U–Pb zircon ages of both sill and dike of the ACC near Lentas are almost the same within uncertainties and suggest emplacement between 72.0 and 73.0 Ma. A monzogranitic vein, which cuts through metasediments of the ACC in the Melambes area (Fig. 2d Supplement), yielded a concordia age at 74.1 ± 0.3 Ma (Martha et al. 2017). Compared to the range of U–Pb zircon ages obtained from the large metagranitoid bodies of the ACC (71–77 Ma, Melambes area, Martha et al. 2017; 72–76 Ma, Pefkos area, Martha et al. 2019, 74–75 Ma, northern Ierapetra Graben, Kneuker et al. 2015), the new U–Pb zircon ages of the igneous veins fit with the younger ages of these bodies, which seems to be reasonable because some of the large granitoid bodies are also cut by igneous veins and are ductility deformed, particularly if quartz is present as a constituent phase. Besides the granitoids, which were emplaced at upper structural levels (e.g., those of the northern Ierapetra Graben, Kneuker et al. 2015), deformation of the large granitoid bodies took place under similarly high temperatures (> 600 °C, Martha et al. 2017, 2019) like those obtained from the igneous veins of the present study. Folding and boudinage of these veins indicate post-emplacement deformation that is younger than the dominant foliation and the pervasive high-temperature deformation of the metasedimentary host rocks. Nevertheless, the deformation microfabrics of the constituent minerals of the veins suggest that folding and boudinage occurred at still high temperature close to the metamorphic peak. Replacement of K-feldspar by myrmekite is evidence for diffusion-assisted deformation (Simpson and Wintsch 1989; Vernon 1991; Ceccato et al. 2018), which occurs at T > 450–500 °C (Cisneros-Lazaro et al. 2019). Partial recrystallization of feldspar suggests T > ca. 500 °C, and chessboard-type subgrains in quartz indicate deformation at T > 600 °C (see references above). Moreover, lobate phase boundaries between quartz and K-feldspar should result from solid-state diffusion creep (e.g., Gower and Simpson 1992; Kruse and Stünitz 1999; Zulauf et al. 2022), which is also consistent with deformation at high temperature. Thus, boudinage of subhorizontal veins (sills) and folding of subvertical veins (dikes) occurred at deeper structural levels under amphibolite facies conditions. Boudinage of subhorizontal competent layers reflects vertical shortening and horizontal extension of the ACC-type rocks. Similar competent sheets, which are affected by boudinage, are subhorizontal quartz veins and calc-silicate layers (Martha et al. 2017). K–Ar ages of amphibole of these ACC rocks are ca. 70 Ma (Seidel et al. 1976). Thus, deformation of the igneous veins should have occurred in Late Cretaceous (Campanian) times between 73 and 70 Ma.

Late Cretaceous arc-type granitoids and Late Cretaceous HT–LP metamorphic host rocks are also exposed further to the north in the Upper Cycladic Unit, and both are commonly correlated with the Pelagonian Zone of continental Greece (Bonneau 1972, 1984; Dürr et al. 1978; Soukis and Papanikolaou 2004; Tortorici et al. 2012; Papanikolaou 2021). A Late Cretaceous low-pressure/high-temperature imprint has been described from Donoussa and Ikaria (Altherr et al. 1994) and from Tinos (Patzak et al. 1994). On Anafi, arc-type granitoid emplacement took place between 69 and 76 Ma (U–Pb on zircon, Reinecke et al. 1982; Martha et al. 2016). Late Cretaceous U–Pb zircon ages have also been determined from metaigneous mélange exposed on Samos and Syros (Tomaschek et al. 2003; Bröcker and Keasling 2006; Bröcker et al. 2014). Late Cretaceous igneous rocks and Late Cretaceous HT–LP metamorphic rocks, however, have not been described from the Pelagonian Zone of continental Greece, which consists of Triassic to Jurassic carbonates resting below late Jurassic to Early Cretaceous (Eo–Hellenic) wildflysch and ophiolite, which are covered by Late Cretaceous transgressive carbonates and a Late Cretaceous to Eocene flysch (Pe-Piper and Piper 2002; Papanikolaou 2021).

It should be emphasized that the Late Cretaceous arc-type rocks of both Crete and the Cyclades are forming or are part of the Uppermost nappes and thus are allochthonous crustal blocks. Their provenance area is still unclear, but should be located further to the north, beyond the Sava–Vardar–Izmir–Ankara Suture, where a continuous Late Cretaceous magmatic arc ranges from the Lesser Caucasus to the Balkans. This domain is almost the same like that where late Permian rift-related igneous rocks are exposed (Fig. 10). The Late Cretaceous arc magmatism in the Black Sea region lasted from 90 to 70 Ma (Okay and Nikishin 2015). The same holds for the Strandja Massif, Srednogorie and Rhodopes, where the U–Pb zircon ages of granitoids range from 92 to 67 Ma (e.g., von Quadt et al. 2005, 2007). Late Cretaceous calc-alkaline magmatic rocks, known as ‘banatites’, are present in the Banatitic Magmatic and Metallogenetic Belt of Romania, Serbia and Bulgaria (e.g., Grünenfelder et al. 1983; Soroiu et al. 1986; Ciobanu et al. 2002; Vlad and Berza 2003). This belt is ca. 1.500 km long and ranges from the Strandja Massif via Sredna Gora and south Carpathians until the Apuseni Mountains (Fig. 10).

The detrital zircon age spectra of the Greenschist Unit and of the ACC are different to those reported above. As the quartzites of both also include late Permian and Mesozoic zircons, which extend until Jurassic and Cretaceous times, their maximum deposition age is much younger than that of the Preveli and Vatos rocks. When taking into account the youngest detrital zircon of the Greenschist Unit (157 ± 4 Ma), its maximum deposition age should be Late Jurassic. Similar to the Vatos quartzite, Cambrian and Silurian detrital zircons are lacking. The Precambrian age spectrum is almost the same like that of the Minoan-type age spectrum described above from the Preveli metaconglomerate. There is clear evidence for late Variscan magmatic activity in the source area, particularly at the Carboniferous/Permian boundary. Such peaks are also present in the metasedimentary rocks of the Tyros Unit (Zulauf et al. 2018). However, in contrast to the latter, the detrital zircon ages of the Greenschist Unit do not cease in the early Permian, but continue via the middle and late Permian until the Late Triassic (Fig. 9). The same holds for the detrital zircon age spectra of the ACC samples, which are similar to those described from the Greenschist Unit above (Fig. 9): (1) The youngest detrital zircons are younger than Triassic, (2) the Precambrian detrital zircons are in line with a Minoan-type age spectrum, (3) all ACC samples include early Paleozoic (Cambro-Ordovician) detrital zircons, (4) Devonian and Carboniferous detrital zircons reflect Variscan magmatism in the source areas, and (5) there is an almost continuous record of detrital zircon ages from Carboniferous to the Triassic or even to the Jurassic or Cretaceous.

The Devonian and Carboniferous detrital zircons of the ACC-type rocks may have come directly from exposed magmatic host rocks or they were recycled from Triassic or Permian metasedimentary rocks of the hinterland. The Triassic detrital zircon ages of the quartzite of the Greenschist Unit and those of the ACC are matching well the U–Pb ages of skarn-grandite and of zircon of felsic metavolcanic rocks of the Preveli Unit (Zulauf et al. 2023a). Moreover, igneous rocks with similar protolith ages have been described from the Cycladic Blueschist Unit (Naxos, Reischmann 1998; Syros, Tomaschek et al. 2001; Andros, Ios and Sifnos, Bröcker and Pidgeon 2007).

To shed light on the source area of the rocks of the Greenschist Unit and of the ACC, we should focus on the presence of late Permian zircons, which are largely lacking in the Tyros Unit of the External Hellenides (Zulauf et al. 2013, 2018), but are present in the Cyclades, where rocks with similar young maximum deposition age are exposed. Detrital zircon U–Pb ages in the Cycladic Blueschist Unit suggest Triassic to Late Cretaceous depositional ages (Löwen et al. 2015; Bröcker et al. 2016; Hinsken et al. 2015; Seman et al. 2017; Flansburg et al. 2019; Poulaki et al. 2019; Laskari et al. 2022). The Cretaceous metasediments of the Cycladic Blueschist Unit include late Permian detrital zircons (Hinsken et al. 2015; Seman et al. 2017; Flansburg et al. 2019; Poulaki et al. 2019; Laskari et al. 2022). The same holds for the Makrotantalon Unit of Andros, where U–Pb dating of detrital zircons yielded ca. 260 Ma (unpublished data cited in Huyskens and Bröcker 2014). In the western Cyclades, flysch-type rocks of the Cycladic Blueschist Unit with Late Jurassic to Cretaceous maximum deposition age and Paleozoic detrital zircons include late Permian detrital zircons (Seman et al. 2017). These flysch-type sediments could be equivalents of the Greenschist Unit and of the ACC rocks of Crete. Late Permian detrital zircons have also been detected in Triassic sedimentary rocks of the Sakarya Zone of NW Turkey (Ustaömer et al. 2016).

However, neither in the Sakarya Zone nor in the Cyclades or in the Pelagonian Zone of mainland Greece, late Permian igneous rocks are exposed. The same holds for Crete. Late Permian (ca. 260 Ma) rift-related granitoids are present further to the north in the Istanbul Zone, Strandja Massif and Dobrogea, most of which extended into Triassic times (Fig. 10; Nikishin et al. 2002; Okay and Nikishin 2015 and references therein; Sałacińska et al. 2021). Late Permian igneous activity is further documented on Lesvos (Koglin et al. 2009b), in the Serbo-Macedonian Zone (Antić et al. 2015, 2016), in the Rhodope (Liati 2005; Bonev et al. 2019), in the Sredna Gora Mountains (Tanatsiev et al. 2012; Lazarova et al. 2021) and in the western Carpathians (Vozárová et al. 2020; Ondrejka et al. 2021; Villaseñor et al. 2021). Finally, similar late Permian U–Pb zircon ages have been derived from A-type rhyolites in the Silicic Unit of Slovakia (Ondrejka et al. 2018), volcanic dikes in the Tatric/Infratatric Units (Putiš et al. 2016; Pelech et al. 2017), rhyodacitic volcanics of the Tisza Mega-Unit, Hungary (Szemerédi et al. 2020) and the Romanian Carpathians (Pană et al. 2002).

As late Permian detrital zircons and Late Cretaceous arc-type granitoids are main constituents of the Greenschist Unit and of the ACC of Crete, their provenance area should be located inside the late Permian/Late Cretaceous igneous belt described above (Fig. 10). The same holds for the Talea Ori group of central Crete. Detrital zircons of the Bali and Sisses Formation of the lower Talea Ori group display middle/late Permian ages (ca. 269 Ma, Zulauf et al. 2016; Seybold et al. 2019), which are compatible with a provenance in NW Turkey as suggested by Seybold et al. (2019). Thus, the Talea Ori group can also be attributed to the Uppermost Unit of Crete.

The role of the North Anatolian Fault Zone

Although the rocks of the Vardar–Serbo–Macedonian–Rhodope domain are similar to those of the Uppermost Unit of Crete, a correlation is questionable because the Late Jurassic/Early Cretaceous blueschists of mainland Greece have been thrust on top of the Pelagonian Zone (Most 2003), which rests on top of Pindos rocks. On Crete, on the other hand, the Uppermost Unit is resting directly on the Pindos nappe, whereas the Pelagonian Zone is lacking. Moreover, the distance between the Vardar–Serbo-Macedonian–Rhodope area and Crete is ca. 700 km, and significant parts of the tectonic nappe transport in northern Greece were accommodated by SW-directed thrusting (Burg 2012, and references therein, Kydonakis et al. 2015) (Fig. 10).

Possible equivalents of the Uppermost Unit of Crete, previously situated in the Pontides or in the southern prolongation of the Strandja Zone, might have been shifted towards Crete during the westward/southwestward counterclockwise movement of the Anatolian block related to dextral strike–slip along the North Anatolian Fault Zone. In NW Turkey the North Anatolian Fault Zone might have initiated as a dominantly N–S-extensional system that later evolved into a strike–slip fault (Le Pichon et al. 2016). Shear-related basins along the Southern Strand are late Miocene in age (Şengör et al. 2005). K–Ar and ⁴⁰Ar/³⁹Ar-dating of authigenic illite (Uysal et al. 2006; Boles et al. 2015) as well as structural and isotopic data of the Eskişhehir and Kapıdağ shear zone (Okay et al. 2008; Türkoĝlu et al. 2016), however, suggest that the dextral movement along the North Anatolian Fault Zone started already in Eocene times, which is consistent with the formation of Eocene pull-apart basins (Ottria et al. 2017). Due to dextral slip along the North Anatolian Fault Zone, slices of the Central Pontide Supercomplex or of similar rocks of the Sakarya Zone or of the Strandja Massif might have been scraped off and translated towards the Pindos basin in the west. Parts of these slices underwent subduction in the Aegean domain (Cyclades), where the main HP–LT metamorphism was initiated at the same Eocene time (e.g., Altherr et al. 1979; Bröcker et al. 1993; Lagos et al. 2007) when dextral slip of the North Anatolian Fault Zone started to be active. The Makrotantalon Unit of Andros might represent such a subducted slice. The Makrotantalon Unit includes marble that yielded Permian fossils (Papanikolaou 1978) and shows evidence for early Cretaceous HP–LT metamorphism (ca. 116 Ma, Huet et al. 2014). As Permian marble and Eohellenic HP–LT metamorphism are also documented in the Preveli Unit of Crete (Bonneau and Lys 1978; Seidel et al. 1977) the latter might form an equivalent of the Makrotantalon Unit.

Thrusting of the Preveli and higher nappes on top of the Pindos Unit in Eocene times was related to top-to-the west movements, which persisted until the early Miocene when the Pindos and Uppermost Unit of Crete were thrust on top of the Tripolitza rocks (Klein et al. 2023). Further movement of these nappes was controlled by slab roll-back and top-to-the south thrusting related to the emplacement of the Tripolitza, Pindos and Uppermost Unit on top of the structurally deeper nappes (Tyros, Pre-Alpine Basement, Phyllite–Quartzite s.str, Typali, Plattenkalk). Such a sequence of nappe emplacement would also explain why fragments of the Late Cretaceous arc—now situated in the Rhodope–Strandja domain and north of the North Anatolian Fault Zone—are major constituents of both the Uppermost Unit of Crete (ACC) and the Upper Cycladic Unit (e.g., Ikaria, Donoussa, Anafi). At present, there is no model and no paleogeographic reconstruction available (e.g., Jolivet and Brun 2008: Fig. 10; van Hinsbergen and Schmid 2012: Fig. 11; Menant et al. 2016), which could explain this peculiar situation.

The U–Pb zircon ages of the reddish granite pebble, that is part of the second Pindos flysch, are consistent with the detrital zircon ages described above. The concordia age at 234 ± 2 Ma is interpreted as the emplacement age of the granite. This emplacement age and the younger inherited zircons (245 ± 6, 246 ± 6) are in line with the age of the felsic volcanic rocks of the Preveli Unit (237 ± 2, 242 ± 1 Ma) and with the Triassic detrital zircons separated from the Greenschist Unit and from the ACC-type rocks. The same holds for the late Permian detrital zircons (262 ± 7 and 271 ± 7 Ma). It should be noted that granitoids with such a Ladinian age have yet not been found in the Pelagonian Zone, but are frequent in the Cyclades, Menderes and Serbo-Macedonian–Rhodope–Strandja domains (Fig. 10). Moreover, the Triassic granite boulder is compatible with slices of Triassic pelagic limestone, described from this nappe contact (Robert and Bonneau 1982). Thus, during the deposition of the Paleocene/Eocene Pindos flysch, the early Triassic granitoids of the Preveli nappe were already exhumed, removed and deposited as olistoliths in the flysch basin. Such a sequence of events is compatible with the Eocene fission-track ages of the Uppermost Unit mentioned above.

Conclusions

Based on the new data and results, the following conclusions can be drawn:

• The detrital zircons of all metasedimentary rocks of the Uppermost Unit of Crete are characterized by a Mionan-type age spectrum and by Carboniferous (Variscan) ages.

• The rocks of the Greenschist Unit and ACC are Jurassic/Cretaceous in age and thus unexpectedly young. Their different degree of LP–HT metamorphism can be explained by Cretaceous slab roll-back, which ‘shifted’ the Late Cretaceous arc towards the accreted trench and fore-arc deposits portrayed by the Vatos rocks.

• The large amount of late Permian detrital zircons together with the presence of Late Cretaceous arc-type granitoids, suggest that the ACC is derived from the late Permian/Late Cretaceous magmatic belt situated north of the Sava–Vardar–Izmir–Ankara Suture in the Strandja–Rhodope domains.

• The distance between the source area of the Uppermost Unit in the Strandja–Rhodope domains and Crete is more than 500 km. The kinematic data suggest that this large distance of nappe movement was accommodated not only by nappe transport, but also by dextral strike–slip along the North Anatolian Fault Zone and related counterclockwise rotation of the Anatolian block.

• There is clear evidence that the Upper Cycladic Unit is an equivalent of the ACC and Greenschist Unit, while the Makrotantalon Unit of Andros might form a Cycladic equivalent of the Preveli nappe. As the rocks of the Upper Cycladic Unit are also allochthonous crustal blocks, they can be regarded as ‘footprints’, which are tracking the movement of the Cretan nappes.

• The olistolith of reddish granite within the second Pindos flysch is of Ladinian age and should be derived from the Preveli Unit, which was situated and eroded in the upper plate during Eocene nappe emplacement.

• Future studies should focus on possible source rocks and on the age of the detrital zircons of the Vatos flysch and of the metasediments, which are part of the nappe contact between Preveli and Pindos Unit.